For the first time since 2021, the Bank of Ghana is no longer in the gold-buying business. In July 2026, the central bank, GoldBod, and the government signed a memorandum of understanding formally transferring Ghana’s Domestic Gold Purchase Programme (DGPP)—and every quasi-fiscal risk that comes with it—entirely to the Ghana Gold Board.
The transfer is being sold, fairly, as a clean-up: a programme that generated estimated losses exceeding US$1.7 billion in 2025 alone is finally off the central bank’s books. But a closer reading of the IMF’s account of the handover, published in its Sixth Review Staff Report and companion Selected Issues Paper over the past six weeks, tells a more complicated story. The Bank of Ghana hasn’t shut the programme down. It has handed it, expanded, to GoldBod—with a demanding list of conditions attached. What was once the central bank’s problem is now entirely GoldBod’s.
From Buying Agent to Principal Trader
Until early 2026, GoldBod’s role in the gold trade was narrow: it acted as a buying agent for the Bank of Ghana, sourcing artisanal gold through a network of aggregators and earning a fee for doing so. The Bank of Ghana financed the purchases, took on the trading and pricing risk, and absorbed whatever losses resulted.
That arrangement is now gone. According to the IMF, GoldBod stopped receiving Bank of Ghana funding for its gold purchases from March 2026. From August 2026, it began raising its own money for artisanal gold purchases directly from commercial banks and off-takers—the exporters who buy the gold from GoldBod once it has been aggregated. The Bank of Ghana’s role has shrunk to that of a fiscal agent: it still handles the foreign exchange mechanics and still receives gold destined for the country’s official reserves, but it no longer owns the risk.
The IMF classifies the transfer as a “prior action”—Fund language for a condition that had to be completed before the IMF would even complete its review of Ghana’s programme. In other words, this was not a voluntary restructuring GoldBod and the Bank of Ghana arrived at on their own timeline. It was a condition the IMF required.
What the Programme Actually Cost
The DGPP’s losses were small at first and grew explosively as it scaled. Total gold purchases under the programme increased from 56.47 tonnes in 2024, valued at approximately US11.4 billion. The IMF estimates that DGPP-related losses exceeded US$1.7 billion in 2025, equivalent to about 17% of the value of doré gold sold by the Bank of Ghana.
The IMF identifies three principal components of those losses:
· Service and assay fees paid to GoldBod
· Discounts on gold sold to off-takers
· Exchange-rate losses arising from the spread between the forex-bureau rate paid to purchase the gold and the cedi reference rate used for Bank of Ghana accounting
Of the three, the IMF says the exchange-rate spread was “most importantly” the biggest single driver, roughly half the total loss on its own. GoldBod’s fees were the smallest slice.
GoldBod CEO Sammy Gyamfi has drawn a clear distinction between GoldBod’s financial performance and the reported losses, arguing that the 0.258% assay fee and 0.5% service fee paid to GoldBod amounted to only 0.758%—compared with the approximately 17% loss reported by the IMF. He has described the losses as largely the result of the policy design of the programme and exchange-rate valuation effects, within a broader economic stabilisation strategy that saw Ghana’s international reserves rise from US13 billion in 2025, alongside a 41% appreciation of the Ghana cedi.
What Actually Changes for GoldBod’s Operations
Four things are fundamentally different for GoldBod now that it no longer operates as a mere buying agent.
- It now owns the trading risk, not just the fee. Previously, GoldBod earned a fee regardless of whether the Bank of Ghana made or lost money on the gold it bought. That protection is gone. GoldBod’s expanded trade model, which began in April 2026, means it now buys, holds, and sells gold on its own account. If the same forex-spread and off-taker-discount losses that hit the central bank recur, they now land directly on GoldBod’s own books.
- It must finance itself. Gold purchases used to be funded with Bank of Ghana money. Since August 2026, GoldBod has been sourcing that financing itself from commercial banks and the same off-takers it sells gold to. Under the new model, commercial banks provide short- and medium-term financing for gold purchases, while off-takers provide funding through structured arrangements linked to the purchase and export of gold. The approach is expected to create a revolving financing cycle in which GoldBod accesses funds to purchase gold, aggregates supplies, and sells to approved off-takers, with proceeds from those transactions supporting subsequent operations.
- Its costs are under a hard target it hasn’t hit yet. The memorandum of understanding formalising the handover sets a cost ceiling of 5% of the value of gold purchased, covering fees, discounts, and the forex spread combined. Real progress was made, cutting costs from roughly 14–15% in 2025 to around 11–12% in early 2026. But that is still more than double the target it is contractually required to reach.
- It now has to publish audited numbers the public can actually check. Under the IMF-backed reform programme, GoldBod must publish externally audited annual financial statements by end-June 2027, with detailed reporting on gold volumes, operational costs, and any transfers it receives from government. This is a formal, dated commitment—not a vague promise of “transparency.”
Already Happening: Buyers Say They’re Not Being Paid
Recent Reuters reports indicate that GoldBod has gone as long as three weeks without paying some of its licensed gold suppliers, forcing some buyers to halt purchases or borrow to stay afloat—this despite gold prices currently surging on international markets.
“The last two weeks were terrible. You could spend a whole day waiting and not receive any funds,” a gold trader in the Ashanti Region told Reuters. A buyer in the Western Region said he had gone unpaid for about three weeks. Five industry sources spoke to the news agency, all on condition of anonymity.
GoldBod disputes that there is any shortfall. The agency said its gold-purchasing operations remained fully funded and operational and rejected any suggestion it had abandoned or was unable to finance its statutory mandate, saying funding decisions are instead based on a supplier’s creditworthiness, security, and risk profile.
But the timing lines up closely with exactly the operational shift identified above. Kwaku Ohemeng Amoah, chief executive of the Chamber of Gold Buyers, told Reuters that GoldBod’s move to fund the trade from its own balance sheet after decoupling from the Bank of Ghana may itself have contributed to the funding constraints and suggested buyers may need to seek supplementary financing of their own.
GoldBod’s CEO, Sammy Gyamfi, has stated that GoldBod raised nearly US75 million before being paused for consultations with the central bank. Three banking executives said the Bank of Ghana viewed GoldBod’s auction programme as inconsistent with its operating framework and that both institutions were working to address the concerns. Fewer than five banks participated in the auction programme, with lenders reportedly more comfortable when the central bank backstopped the arrangement.
A Bigger Bet on a Single, Volatile Commodity
At the same time GoldBod is being asked to cut costs and absorb full risk, the government wants it to buy considerably more gold than the Bank of Ghana ever did. Under the Ghana Accelerated National Reserve Accumulation Policy (GANRAP), the plan is to push reserves to 15 months of import cover by 2028, well beyond the roughly 6 months the IMF itself considers adequate.
GoldBod is central to how that happens. The programme targets at least 8.6 months of import cover by the end of 2026, more than 11.8 months by the end of 2027, before reaching the 15-month target in 2028. Government has set an operational weekly gold purchase target of approximately 3.02 tonnes, achieved through acquisition of at least 2.45 tonnes weekly from the artisanal and small-scale mining sector and invocation of pre-emption rights to secure a minimum of 0.57 tonnes weekly from the large-scale mining sector.
Under GANRAP, GoldBod is expected to:
· Buy twice as much large-scale mine gold as before, targeting 30% of total large-scale output, up from around 20%
· Refine gold domestically rather than importing it already refined, then validate it at a newly designated LBMA-certified refinery in South Africa with new infrastructure and logistics GoldBod didn’t need to manage before
· Buy slightly more artisanal gold over the next three years than it did even in 2025, the year that produced the massive losses
· Run a bigger aggregator network: GoldBod has already licensed a second aggregator alongside Bawa-Rock, which alone supplied over 60% of all artisanal gold exported in 2025
The IMF is blunt about what this will cost. At 15 months of import cover, the expense of “sterilising” the extra reserves—the routine central-bank operations needed to mop up the cedi liquidity created whenever gold or foreign exchange is bought—would likely reach 3% of GDP, up from about 1% in 2025. That is a cost separate from any trading loss, and it lands on the wider economy, not just GoldBod’s balance sheet.
The Risks GoldBod Now Carries Alone
Handing the DGPP to GoldBod solved one specific problem the IMF had flagged: the mismatch where one institution earned fees while another absorbed the losses, what the Fund itself called a “moral hazard problem.” That mismatch is technically fixed now. GoldBod earns the fees and bears the losses. But that fix creates a new, more concentrated set of risks.
Can GoldBod actually hit the cost target? Getting costs from today’s roughly 11–12% down to the mandated 5% requires structural changes that GoldBod hasn’t fully delivered yet—reducing the forex spread, streamlining a supply chain that runs through 13 mining regions, and negotiating better terms with off-takers who have historically had the upper hand in doré gold deals. If GoldBod can’t get there, it now absorbs the shortfall itself, without a central bank balance sheet to quietly cushion it.
Is its financial cushion as strong as it looks? GoldBod’s audited 2025 accounts show a surplus of GH₵5.44 billion, reassuring on its face. But GH₵4.55 billion of that is an unspent government capital subvention, not trading profit. Strip that out, and GoldBod’s actual operating surplus from its fee-based business was GH₵909.7 million. That is the more realistic number to weigh against the scale of what GoldBod is now being asked to finance: a doubled large-scale gold purchase target, a new refining operation, and a compliance overhaul—all without government money doing the heavy lifting a second time. The Reuters reporting suggests this gap between headline balance-sheet strength and actual day-to-day liquidity may already be showing up on the ground.
A bigger bet on a single, volatile commodity. Gold now makes up more than half of Ghana’s total exports, up from a fifth in 2021. The IMF’s own debt sustainability analysis tests what happens if global gold prices fall 30%: it would force a materially tighter budget, with the government’s primary balance target rising from 0.5 to 0.8% of GDP just to keep debt on track. The more Ghana leans on gold to build reserves, the more exposed the whole economy becomes to a single commodity price GoldBod does not control.
Bright Simons, Vice President of IMANI Africa, has renewed concerns over the sustainability of GoldBod’s model, arguing that the Bank of Ghana incurred an average cost margin of about 17 cents for every dollar of foreign exchange generated through GoldBod in 2025. While acknowledging the government’s position that costs have fallen to below 12 cents per dollar with intentions to reduce further to below five cents, he noted that those calculations did not include the cost of “sterilisation”. He also questioned whether the “benefits” of the programme—including a more stable cedi and lower inflation—were due mostly to the hike in gold prices and “would have manifested anyway without the GoldBod”.
Smuggling that formalisation still hasn’t beaten. Despite GoldBod’s mandate to formalise the sector, the IMF estimates 229 tonnes of gold worth US4 billion. Every tonne that leaves informally is a tonne GoldBod’s reserve-building strategy can’t count.
Environmental costs the IMF says are becoming structural. The same artisanal mining expansion that feeds GoldBod’s gold supply is, in the IMF’s own words, degrading cocoa farmland, silting
